Selection for Urease Activity during Helicobacter pylori Infection of Rhesus Macaques (Macaca mulatta)

Abstract

Helicobacter pylori strain J166 recovered from experimentally inoculated rhesus monkeys had up to a 250-fold-increased urease activity over that before inoculation. This was found to result from the selection of urease positive J166 clones from a heterogenous inoculum, which was predominantly urease negative due to a 1-bp insertion in the ureA gene. These results confirm the importance of urease for H. pylori colonization. Strain J166 is particularly well adapted to the rhesus monkey, since it colonized preferentially despite the fact that less than 0.1% of the inoculum was urease positive.

The rhesus monkey has been studied (5–8, 14, 15) as an animal model of Helicobacter pylori pathogenesis. We sought to confirm previous work that showed that H. pylori strain J166 is well adapted to colonization of the rhesus monkey (5). A mixture of H. pylori J166 and two other H. pylori strains was inoculated into monkeys, and, as expected, strain J166 predominated in all three animals. Surprisingly, the urease activity of H. pylori J166 increased markedly from pre- to postinoculation, which was of interest because of the well-known importance of urease in H. pylori pathogenesis. Here we provide a molecular characterization of this phenomenon and discuss its implication for our understanding of H. pylori in the rhesus model.

Animal procedures.

Three male and female rhesus macaques aged 1 to 2 years documented to be free of H. pylori and “Helicobacter heilmannii” (15) were inoculated with human-derived H. pylori strains D5127, 88–23, and J166, the last of which has been shown to persistently colonize when experimentally inoculated into rhesus monkeys (5). Each strain could easily be identified by agarose electrophoresis of repetitive extragenic palindromic PCR (Rep-PCR) products (10). Monkeys were inoculated with a mixture of approximately 10⁹ bacteria containing equal numbers of each strain grown in liquid culture (15). Four weeks following inoculation, gastric mucosal biopsy samples were obtained by endoscopy and H. pylori was cultivated as previously described (15). Examination of between four and nine colonies from each monkey by Rep-PCR showed that only strain J166 was recovered from all three animals (data not shown).

H. pylori urease assays.

Prior to inoculation, the three H. pylori strains were examined by a qualitative rapid urease assay with indole-urea medium (11), which confirms the presence of urease by a color change resulting from hydrolysis of urea and a rise in pH. Surprisingly, while strains D5127 and 88–23 displayed the typical rapid urease reaction of color change in less than 5 min, the inoculated J166 exhibited color change only after 2 to 3 h. However, all H. pylori colonies recovered from the three monkeys showed the rapid color change that is typical of H. pylori, despite the fact that they were each identified as strain J166 by Rep-PCR. The qualitative difference in urease activity between the inoculated J166 and the recovered J166 was confirmed quantitatively using a modification of the Berthelot reaction (3). Five replicate assays were performed on each isolate. Urease activity for the recovered H. pylori J166 ranged from 100- to 250-fold higher than that of the inoculated J166 (Table 1).

TABLE 1.

Urease genotypes and activities of strains used in this study

Strain	Urease genotype	Urease activity (mean ± SD) (nmol-urea/min/mg of protein)
H. pylori
J166	Wild type (inoculated)	79.0 ± 4.0
	Wild type (recovered; monkey 1)	20,280.0 ± 1,350.0
	Wild type (recovered; monkey 2)	7,902.0 ± 1,010.0
	Wild type (recovered; monkey 3)	13,018.0 ± 720.0
E. coli
SE5000(pJ116)	ureABIEFGH nixA (inoculated)	0.4 ± 0.8
SE5000(pJ105)	ureABIEFGH nixA (recovered)	30.0 ± 1.4
SE5000(pJ116)  (pJ121)	ureABIEFGH nixA (inoculated), ureAB (recovered)	35.0 ± 2.3
SE5000(pJ116)  (pJ126)	ureABIEFGH nixA (inoculated), ureAB (inoculated)	0.01 ± 0.4

Cloning and urease activity in Escherichia coli.

To determine the molecular basis for the increased urease activity in the recovered J166, we cloned the urease operon (ureABIEFGH) and the nixA gene, which is necessary for full urease activity (12), from either the recovered (pJ105) or the inoculated (pJ116) H. pylori J166. Template DNA used for construction of pJ105 was obtained from H. pylori J166 recovered from monkey 3 (Table 1). Amplification of ureABIEFGH and nixA was performed using primers (Table 2) derived from published sequences (4, 11, 12) and standard PCR conditions. The amplified products were ligated into pACYC184 (ATCC 37033) and transformed into E. coli SE5000 (provided by H. Mobley). Quantitative urease assays were performed as described above, with the mean background value for E. coli SE5000 subtracted from each result. A 75-fold increase in urease activity was observed for E. coli SE5000(pJ105) containing the urease operon and nixA gene of the recovered H. pylori J166, compared to E. coli SE5000(pJ116) with the urease operon and nixA gene from the inoculated H. pylori J166 (Table 1). These results suggested that the difference in urease between inoculated and recovered J166 was due to genetic alteration either in the urease operon or in nixA. To further localize the faulty gene, constructs were made combining the urease operon from the inoculated J166 with the nixA gene from the recovered J166 and vice versa. Qualitative urease assays were performed on E. coli SE5000 transformed with these recombinant plasmids. After overnight incubation, urease activity was detected only in the construct containing the urease operon from the recovered J166, which indicated that the urease operon from the inoculated J166 was defective.

TABLE 2.

PCR primers used in this study

Amplicon	Forward primer (5′ to 3′)	Reverse primer (5′ to 3′)
ureABIEFGH	CCATGCGATAGAGTTTGGCATGGTG	CAAGTTAAAAGAATTGTTAAACAACCACTTC
nixA	GTATGCCTATTTAGAAGCGCTTGGATTAG	TGCCAGTTTTTAGCCTCCCCTTTGTTTG
ureAB	CCATGCGATAGAGTTTGGCATGGTG	GTGCTTTTAGGATCGACTTTGG
ureIEF	TACCATGTGTTCGTGGATGG	GTTTTACCGCTTCCTACAGG
ureGH	AAGGCGATGCAGCATGAGAG	CAAGTTAAAAGAATTGTTAAACAACCACTTC

Complementation and sequencing.

Complementation experiments were then performed to further localize, within the urease operon, the genetic change responsible for the increased urease activity in the recovered H. pylori J166. We PCR amplified ureAB, ureIEF, and ureGH from the recovered J166 (from monkey 3) using primers shown in Table 2. Each fragment was ligated into pBluescript (Life Technologies GIBCO BRL, Gaithersburg, Md.) and transformed into E. coli SE5000(pJ116). Each E. coli construct was tested qualitatively for urease activity. After overnight incubation, urease activity was detected only when complementation was performed with pJ121, which contained ureAB from the recovered J166. When measured quantitatively, urease activity was markedly increased and was similar to that of E. coli SE5000(pJ105) (Table 1). To be sure that the increased copy number of the urease structural subunits expressed from pJ121 was not responsible for the increased urease activity, we also complemented E. coli SE5000(pJ116) with ureAB from the inoculated J166 (pJ126). No increase in urease activity was observed (Table 1). These results indicated that the increased urease activity of the recovered H. pylori J166 was due to an alteration in the genes coding for the urease structural subunits, ureAB.

The ureAB genes from the inoculated J166 and from one recovered J166 strain (from monkey 3) were amplified and completely sequenced using an ABI Prism 377 automated DNA sequencer and the ABI Prism dRhodamine Terminator Cycle Sequencing kit (PE Biosystems, Foster City, Calif.). The sequences were identical except for a one-base insertion 18 bp after the start of ureA in the inoculated strain (Fig. 1). This insertion introduced a stop codon at a distance corresponding to 20 amino acids downstream from the ureA start site. The sequences of this region were found to be identical in J166 isolates recovered from all three monkeys.

FIG. 1.

DNA sequence and deduced amino acid sequence of the first portion of the ureA gene from the inoculated and recovered H. pylori J166. The inserted C at 18 bp in the inoculated J166 is shown in bold. The asterisk in the amino acid sequence of the inoculated strain indicates a stop codon.

Western blot.

To confirm the absence of UreA in the inoculated J166, whole-cell lysates were run on a sodium dodecyl sulfate–12% polyacrylamide minigel (5 μg or protein per well), transferred to nitrocellulose, and reacted with polyclonal antiserum specific for UreA (kindly provided by H. Mobley). Bound antibodies were visualized with the ECL kit (Amersham Pharmacia Biotech, Piscataway, N.J.). Urease expression was apparent in the recovered H. pylori J166 and in E. coli SE5000(pJ105) (Fig. 2). Little or no expression of ureA was seen in the inoculated J166 or in E. coli SE5000(pJ116).

FIG. 2.

Immunoblot of H. pylori J166 and recombinant E. coli SE5000 whole-cell lysates reacted with anti-UreA antibody. Lane A, inoculated J166; lane B, recovered J166; lane C, E. coli SE5000 with ureABIEFGH and nixA from inoculated H. pylori J166; lane D, E. coli SE5000 with ureABIEFGH and nixa from recovered H. pylori J166.

Acid selection.

Since the inoculated J166 had some urease activity but appeared to have a UreA protein that was truncated after only 20 amino acids, we considered the possibility that the inoculum was a heterogeneous mixture of urease-positive and urease-negative organisms. To look for urease-positive organisms, we took advantage of the observation that urease-negative H. pylori organisms are killed at low pH in the presence of urea, while urease-positive H. pylori organisms survive due to alkalization of the media from urea hydrolysis (2). Growth from 48-h. plates of inoculated H. pylori J166 was harvested into phosphate-buffered saline (pH 7.2) and adjusted to an A₆₀₀ of 1.0. Cells were pelleted, resuspended in phosphate-buffered saline (pH 1.5) containing 5 mM urea, and incubated at 37°C for 15 min. Serial dilutions were plated before and after acid treatment (Table 3). All cells recovered following acid treatment had an active urease by qualitative assay. We next used the acid tolerance assay to obtain a semiquantitative estimate of the proportion of inoculated J166 that was urease positive. The assay was performed on a urease-positive colony (that survived acid treatment), a urease-negative colony (obtained by plating serial dilutions of the inoculated J166 and testing isolated colonies for urease activity), and the heterogeneous inoculated J166. Based on the results of this assay (Table 3), we estimate that the inoculated J166 consisted of approximately 0.009% urease-positive clones (1.55/1.22 × 10⁻⁴ = 1.27 × 10⁴ urease-positive clones in the pre-acid inoculum of 1.38 × 10⁸ = 0.009%).

TABLE 3.

Effect of acid treatment on survival of H. pylori J166

Isolate	Viable colonies (CFU/ml)		Survival frequency (mean ± SD)
	Before acid	After acid (mean ± SD)
Inoculated J166	1.38 × 108	1.55 ± 1.86	1.13 × 10−8 ± 1.3 × 10−8
Urease-negative J166	1.35 × 108	0.00	0.00
Urease-positive J166	7.90 × 107	9.60 × 103 ± 6.0 × 103	1.22 × 10−4 ± 7.6 × 10−5

It is unknown how this heterogeneous culture of H. pylori J166 arose. It was obtained as a low-passage human isolate that was presumably urease positive and may have contained a small urease-negative population that was enriched in the laboratory. Naturally occurring urease-negative isolates have been described (13), though the frequency with which they occur is unknown. We have been able to recover urease-negative organisms with the same base insertion in ureA from another vial of low-passage J166 from the original source.

These results have three implications for the rhesus monkey model of H. pylori. First, urease-negative strains of H. pylori do not colonize the rhesus monkey. This is not surprising, since the same result has been found in the pig (9) and mouse (16) models of H. pylori and in the ferret model of Helicobacter mustelae (1). Second, these results permit us to estimate the minimum infectious dose of H. pylori J166 in the rhesus monkey, which has not previously been examined. Semiquantitative estimates based on urease assay and acid tolerance suggest that between 1 in 10² and 1 in 10⁴ of the inoculated J166 organisms were urease positive. Since the original mixed inoculum was approximately 10⁹, of which one-third was J166, the minimum infectious dose of J166 in the rhesus monkey is probably less than 10⁶ CFU and perhaps as low as 10⁴ CFU. Finally, these observations provide further evidence for the presence of variation in naturally occurring H. pylori populations and imply that the host selects for the most fit members of the population. H. pylori J166 is particularly well adapted to the rhesus monkey, since urease-positive J166 competed effectively against strains D5127 and 88–23, despite being present in much smaller numbers in the original mixed inoculum.